The recent experimental observation of a metal-insulator transition in two dimensions prompts a re-examination of the theory of disordered interacting systems. We argue that the existing theory permits the existence of a metallic phase and propose a number of experiments such as magnetoconductance and tunnelling in the presence of a parallel field, which should provide diagnostic tests as to whether a given experimental system is in fact in this regime. We also comment on a generic flow diagram which predicts a maximum metallic resistivity.